Scare' Em or Disgust 'Em: The Effects of Graphic Health Promotion Messages

Abstract

This study experimentally tested the effects of 2 types of content commonly found in anti-tobacco television messages—content focused on communicating a health threat about tobacco use (fear) and content containing disgust-related images—on how viewers processed these messages. In a 2 × 2 within-subjects experiment, participants watched anti-tobacco television ads that varied in the amount of fear and disgust content. The results of this study suggest that both fear and disgust content in anti-tobacco television ads have significant effects on resources allocated to encoding the messages and on recognition memory. Heart-rate data indicated that putting fear- or disgust-related content into anti-tobacco ads led to more resources allocated to encoding compared to messages without either feature. However, participants appeared to allocate fewer resources to encoding during exposure to messages featuring both fear and disgust content. Recognition was most accurate for messages that had either fear or disgust content but was significantly impaired when these 2 message attributes occurred together. The results are discussed in the context of motivated processing and recommendations about message construction are offered to campaign designers.

ACKNOWLEDGMENTS

The authors would like to acknowledge the substantial contributions of the Editor and two anonymous reviewers to this article.

Notes

¹The public service announcements we used in our study are commonly referred to as fear appeals. Perhaps a better way to think about them is as a threat appeal. The term fear expresses an emotion felt by an individual, a reaction from an audience member. Threat however, refers to a message attribute, and may or may not result in fear (O'Keefe, 2003; Witte, 1992). We adopt the term fear appeal to label our messages in order to be consistent with the bulk of the literature.

²It was discovered that one person pressed the “no” button for all recognition items, except one. The false alarm rate was 0%, but the hit rates were also 0% across all conditions, except for one item. Two other participants had extremely low hit rates across conditions (ranging from 0% to 17%).